Dynamic Wind Tunnel Load and Attitude Measurements Using State Estimation

Abstract

Sting-mounted models that undergo significant unsteady motion suffer from degraded data quality because the data is time-averaged to remove the unsteady fluctuations. However, time-averaged data is not always an accurate representation of the true data. Eliminating such errors is addressed in this study by developing and evaluating the performance of a Kalman filter for estimating instantaneous load and model attitude data for a sting-mounted wind tunnel model. The particular model is 6.25% scale WB-57 that is tested in the Oran W. Nicks Low Speed Wind Tunnel at Texas A&M University. The pitch and plunge motion of the model are measured using accelerometers and the loads and moments are measured using an internal balance.

This work shows that a simplified state-space model consisting of 3 state variables and one measurement can successfully estimate plunge position and normal force of a sting-mounted test article by minimizing the difference between actual and predicted measurements in the Kalman filter. The aerodynamic normal force results compared well with conventional time-averaged wind tunnel data used as a metric to measure the successfulness of the state estimation technique.

A more extensive state space model with 6 state variables and 4 measurements has the potential to estimate the pitch position and pitching moment in conjunction with the plunge position and loads. Doing so would require a different technique to quantify and tune the process noise covariance matrix.